An interactive explainer
Where transmission is seasonal, most of the year's malaria falls in a few months, so it can pay to time control to the season: spraying houses with insecticide (IRS) before transmission climbs, or giving children a course of preventive drugs through the peak. This explainer follows the second of these, seasonal malaria chemoprevention (SMC).1 Its protection lasts only a short time, so how much disease it prevents depends on when the rounds land and on who they reach.
In much of the Sahel and other highly seasonal settings, rainfall is confined to a short wet season, and malaria broadly tracks it: mosquito numbers, transmission and childhood cases tend to rise and fall over the year, so a large share of the disease in young children arrives in a window of two to four months.2 The more concentrated the season, the more of the annual burden sits inside that window.
Broadly, rainfall creates the standing water that mosquito larvae need, so mosquito numbers, and the rate of infectious bites, tend to build after the rains. The link is loose rather than mechanical: the type of water body, temperature, drainage and how people are exposed all matter, and some settings peak at odd times or stay high year-round. Where the seasonal pattern does hold, transmission and then cases lag the rainfall by weeks, because larvae take time to mature and the adult population builds gradually, and infections then take further time to build and progress to illness,3 giving a single, delayed peak in childhood cases rather than a flat year-round level.
SMC gives young children (commonly those aged 3 to 59 months, though the eligible range varies by setting) a full antimalarial course repeated every month through the transmission season.1 Each course clears any infection the child already has and then guards against new infection while the drug lasts, but only while it lasts: as the drug clears from the body over the following weeks, that protection fades.
We can capture this by letting a dose's protection start high and then decay over time.4 What matters is the shape, near-full protection at first and then a fall, and the duration; together they set how far apart the rounds need to be. Where resistance to the drugs is high the protection is shorter-lived, and alternative regimens are being studied.
Protection from a single dose, against time since it was given (the SP-AQ prophylaxis curve from malariasimulation). It holds high for a short period, then falls away fairly steeply. SP-AQ, sulfadoxine-pyrimethamine plus amodiaquine, is the standard SMC drug combination, and its protective duration is what sets the roughly monthly spacing of rounds.
Because each dose protects for only a short time, a single round cannot cover a season that lasts several months. SMC is delivered as a block of monthly rounds, typically four or five, timed to span the period of highest transmission.2 How much disease the programme prevents depends on how well that block of protection lines up with the peak in cases.
Total protection through the year is the envelope of the individual dose curves, one per round. The number of rounds and the spacing between them set how much of the season is covered:
Because cases are concentrated near the peak, moving the same rounds a few weeks earlier or later changes how much disease they prevent by much more than the change in timing might suggest.
If SMC rounds are scheduled to a fixed calendar set in advance, often from the historical timing of the rains, then when the season arrives earlier or later than that calendar assumes the same rounds line up poorly with the peak and prevent fewer cases. The onset and length of rainy seasons already vary from year to year, and climate change is projected to shift the timing, length and geography of transmission seasons,5 so a calendar fixed to a long-run average can drift out of step with the season that actually occurs.
Shifting the season moves the case peak along the year while the scheduled rounds stay put. The block of protection then straddles the shoulder of the season rather than its peak, so it overlaps fewer cases and averts less. The effect is strongest in the most concentrated seasons, where the peak is narrow and a modest shift can move most of the cases outside the covered window. As the climate changes, seasons may become not only displaced but less predictable from one year to the next, so a calendar built on the past becomes a weaker guide to the season ahead and well-timed rounds harder to plan. Timing the rounds to the observed case peak, rather than to the first rains, keeps them aligned.3
A toy model of childhood cases with and without SMC
Set how seasonal transmission is, when cases peak and how fast they take off into the peak, then choose the SMC schedule (number of rounds, spacing and first round), the coverage per round, and who is reached across rounds: the same children each time, a random draw, or deliberately different children. The chart shows childhood cases through the year without SMC (blue) and with it (purple); the shaded area is the cases averted.
Two things to try: shift the case peak while leaving the rounds fixed, to see a fixed calendar fall out of step; and switch who is reached, to compare covering the same, random or different children. The slider sets the peak in cases, which in reality lags the rains by a few weeks; the toy places that case peak directly rather than simulating the rainfall and the mosquito lag.
Childhood cases through the year
Each SMC round (orange marker) starts a fresh protection curve that then decays, so cases fall furthest just after a round and recover in the gaps between rounds. The rounds prevent the most disease when their combined protection sits over the case peak. Move the peak away from the rounds, or widen the spacing, and the averted area shrinks.
Protection is a scalloped envelope, not a flat shield. Because each dose decays, cases with SMC dip after every round and then climb part-way back before the next one arrives. Spacing the rounds about a month apart, close enough that protection has not fully waned before the next dose, keeps that climb-back small; wider spacing lets cases rise further between rounds.
Alignment with the peak matters more than the number of doses. A block of rounds centred on the case peak averts far more than the same rounds sitting on the shoulder of the season. Sweep the first-round slider with everything else fixed: the averted figure rises to a maximum when the rounds cover the peak and falls away on either side.
The penalty for mistiming grows with seasonality. Raise the seasonality slider so the peak is narrow, then shift the peak while leaving the rounds fixed. A concentrated season loses more when the calendar is off, because most of its cases sit in a short window that the rounds now miss. A flatter season is more forgiving.
Time to the cases, not the first rains. Cases usually lag the rains, and the link between the two is loose, so rounds pinned to the onset of the rains can lead the case peak and cover less than expected. Timing to the observed peak in cases, where it can be measured, is a surer guide.3
Who is reached matters as much as how many, and it is an equity question. At a fixed coverage per round, switch between reaching the same, random and different children. Reaching the same children each round protects a fixed group but leaves the rest exposed all season, so it averts the least; spreading rounds across more children (watch "children given ≥1 round" climb) usually averts more. How coverage correlates between rounds is a model parameter, not just a delivery detail: in real programmes it is positively correlated, the same hard-to-reach children being missed round after round, so the aim is higher coverage and weaker correlation, drawing in the never-covered children rather than deliberately rotating who is protected.6
Methods. A deliberately simple, illustrative toy model showing the shape of how SMC impact depends on timing, not a specific setting; not to be used for decision making. Cases are shown in relative units on an arbitrary scale, so only the shape and the relative reductions are meaningful.
References.